提出了一种基于检测和验证相结合的指针综合提取方案，目的是实现特种车辆指针式仪表图像识别的自动识别和自动矫正。该方案优化利用了 Hough变换法、极坐标中心投影以及边缘点集合的最小二乘法等算法，使得指针的自动提取无论针对何种品质的图像都非常准确。经验证，采用该方法进行特种车辆仪表指针提取，准确度达到 99%以上。

A detection and verification based approach is applied to pointer exact extraction for automatical calibration and automatic recognition of dial instruments for vehicle car. The syntheticapproach optimizes algorithms such as Hough Transformation, polar central projection，Least Squarealgorithm，etc.，which brings the high performance to dial reading results. It is tested that, the veracity of the synthetic approach can reach above 99%.